You are on page 1of 75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Nonlinear Control
Lecture 9: Feedback Linearization
Farzaneh Abdollahi

Department of Electrical Engineering


Amirkabir University of Technology

Fall 2011
Farzaneh Abdollahi

Nonlinear Control

Lecture 9

1/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Feedback Linearzation
Input-State Linearization
Input-Output Linearization
Internal Dynamics of Linear Systems
Zero-Dynamics

Preliminary Mathematics
Diffeomorphism

Frobenius Theorem
Input-State Linearization
Control Design
Input-Output Linearization
Well Defined Relative Degree
Undefined Relative Degree
Normal Form
Zero-Dynamics
Local Asymptotic Stabilization
Global Asymptotic Stabilization
Tracking Control
Farzaneh Abdollahi

Nonlinear Control

Lecture 9

2/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Feedback Linearzation
I

The main idea is: algebraically


transform a nonlinear system dynamics
into a (fully or partly) linear one, so that
linear control techniques can be applied.

In its simplest form, feedback linearization


cancels the nonlinearities in a nonlinear
system so that the closed-loop dynamics
is in a linear form.
Example: Controlling the fluid level in a
tank

Objective: controlling of the level h of


fluid in a tank to a specified level hd ,
using control input u
the initial level is h0 .

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

3/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

The dynamics:

A(h)h(t)
=ua

I
I

I
I

where A(h) is the cross section of the tank and a is the cross section of
the outlet pipe.

Choose u = a 2gh + A(h)v h = v


Choose the equivalent input v : v = h where h = h(t) hd is error
level, a pos. const.
resulting closed-loop dynamics: h + h = 0 h 0 as t

The actual input flow: u = a 2gh + A(h)h


I
I

p
2gh

First term provides output flow a 2gh


Second term raises the fluid level according to the desired linear dynamics

If hd is time-varying: v = h d (t) h
I

h 0 as t

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

4/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Canceling the nonlinearities and imposing a desired linear dynamics, can


be simply applied to a class of nonlinear systems, so-called companion
form, or controllability canonical form:

A system in companion form:


x (n) (t) = f (x) + b(x)u
I
I
I
I

(1)

u is the scalar control input


x is the scalar output;x = [x, x,
..., x (n1) ] is the state vector.
f (x) and b(x) are nonlinear functions of the states.
no derivative of input u presents.

(1)can be
presented
as controllability
canonical form

x1
x2
..

..

d
.
.
=

dt

xn1
xn
xn
f (x) + b(x)u

for nonzero b, define control input: u = b1 [v f ]

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

5/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Feedback Linearzation
I

the control law:


v = k0 x k1 x . . . kn1 x (n1)

ki is chosen s.t. the roots of s n + kn1 s n1 + . . . + k0 are strictly in LHP.

Thus: x (n) + kn1 x (n1) + . . . + k0 = 0 is e.s.

For tracking desired output xd , the control law is:


(n)

v = xd k0 e k1 e . . . kn1 e (n1)
I

Exponentially convergent tracking, e = x xd 0.

This method is extendable when the scalar x was replaced by a vector


and the scalar b by an invertible square matrix.

When u is replaced by an invertible function g (u) u = g 1 ( b1 [v f ]),

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

6/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example: Feedback Linearization of a Two-link Robot


I

A two-link robot: each joint equipped


with
I
I
I

a motor for providing input torque


an encoder for measuring joint position
a tachometer for measuring joint velocity

objective: the joint positions ql and q2


follow desired position histories qdl (t) and
qd2 (t)

For example when a robot manipulator is


required to move along a specified path,
e.g., to draw circles.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

7/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Using
the Lagrangian
the robotic dynamics


 equations


 are:


H11
H21

H12
H22

q1
q2

hq 2
hq 1

hq 2 hq 1
0

q 1
q 2

g1
g2


=

1
2

where q = [q1 q2 ]T : the two joint angles, = [1 2 ]T : the joint inputs, and
2
2
H11 = m1 lc1
+ l1 + m2 [l12 + lc2
+ 2l1 lc2 cos q2 ] + I2
H22

2
2
= m2 lc2
+ I2 H12 = H21 = m2 l1 lc2 cos q2 + m2 lc2
+ I2

g1

= m1 lc1 cos q1 + m2 g [lc2 cos(q1 + q2 ) + l1 cos q1 ]

g2

= m2 lc2 g cos(q1 + q2 ), h = m2 l1 lc2 sin q2

Control law for tracking, (computed torque):



 

 

 

1
H11 H12
v1
hq 2 hq 2 hq 1
q 1
g1
=
+
+
2
H21 H22
v2
hq 1
0
q 2
g2

where v = qd 2q 2 q, q = q qd : position tracking error, : pos. const.


I q
d + 2q + 2 q = 0 where q converge to zero exponentially.
I

This method is applicable for arbitrary # of links

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

8/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-State Linearization
I

When the nonlinear dynamics is not in a controllability canonical form,


use algebraic transformations

Consider the SISO system


x = f (x, u)

In input-state linearization technique:


1. finds a state transformation z = z(x) and an input transformation
u = u(x, v ) s.t. the nonlinear system dynamics is transformed into
z = Az + bv
2. Use standard linear techniques (such as pole placement) to design v .

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

9/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example:
I

Consider

x 1 = 2x1 + ax2 + sin x1


x 2 = x2 cos x1 + u cos(2x1 )

I
I
I

Equ. pt. (0, 0)


The nonlinearity cannot be directly canceled by the control input u
Define a new set of variables:
z1 = x1
z2
z 1
z 2

I
I
I
I

= ax2 + sin x1
= 2z1 + z2
= 2z1 cos z1 + cos z1 sin z1 + au cos(2z1 )

The Equ. pt. is still (0, 0).


1
(v cos z1 sin z1 + 2z1 cos z1 )
The control law: u = a cos(2z
1)
The new dynamics is linear and controllable: z 1 = 2z1 + z2 , z 2 = v
By proper choice of feedback gains k1 and k2 in v = k1 z1 k2 z2 , place
the poles properly.
Both z1 and z2 converge to zero,
the original state x converges to zero

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

10/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

The result is not global.


I

The result is not valid when xl = (/4 k/2), k = 0, 1, 2, ...

The input-state linearization is achieved by a combination of a state


transformation and an input transformation with state feedback used in
both.
To implement the control law, the new states (z1 , z2 ) must be available.
I

If they are not physically meaningful or measurable, they should be


computed by measurable original state x.

If there is uncertainty in the model, e.g., on a


of new state z as well as control input u.

For tracking control, the desired motion needs to be expressed in terms of


the new state vector.
Two questions arise for more generalizations which will be answered in
next lectures:

I
I

error in the computation

What classes of nonlinear systems can be transformed into linear systems?


How to find the proper transformations for those which can?

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

11/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-Output Linearization
I

Consider
x

= f (x, u)

= h(x)

Objective: tracking a desired trajectory yd (t), while keeping the whole


state bounded

yd (t) and its time derivatives up to a sufficiently high order are known
and bounded.
The difficulty: output y is only indirectly related to the input u

it is not easy to see how the input u can be designed to control the
tracking behavior of the output y .

Input-output linearization approach:


1. Generating a linear input-output relation
2. Formulating a controller based on linear control

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

12/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example:
I

Consider

x 1 = sin x2 + (x2 + 1)x3


x 2 = x15 + x3
x 3 = x12 + u
y

= x1

To generate a direct relationship between the output y and the input u,


differentiate the output y = x 1 = sin x2 + (x2 + 1)x3

No direct relationship
differentiate again: y = (x2 + 1)u + f (x), where
5
f (x) = (x1 + x3 )(x3 + cos x2 ) + (x2 + 1)x12

Control input law: u =

Choose v = yd k1 e k2 e,
where e = y yd is tracking error, k1 and k2
are pos. const.

The closed-loop system: e + k2 e + k1 e = 0

e.s. of tracking error

Farzaneh Abdollahi

1
x2 +1 (v

f ).

Nonlinear Control

Lecture 9

13/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

The control law is defined everywhere except at singularity points s.t.


x2 = 1

To implement the control law, full state measurement is necessary,


because the computations of both the derivative y and the input
transformation need the value of x.
If the output of a system should be differentiated r times to generate an
explicit relation between y and u, the system is said to have relative
degree r .

For linear systems this terminology expressed as # poles # zeros.

For any controllable system of order n, by taking at most n


differentiations, the control input will appear to any output, i.e., r n.
I

If the control input never appears after more than n differentiations, the
system would not be controllable.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

14/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Feedback Linearzation
I

Internal dynamics: a part of dynamics which is unobservable in the


input-output linearization.
I

1
yd (t)
x2 +1 (

k1 e k2 e + f )

The desired performance of the control based on the reduced-order model


depends on the stability of the internal dynamics.
I

In the example it can be x 3 = x12 +

stability in BIBO sense

Example: Consider

x 1
x 2


=


x23 + u
u
y = x1

(2)

Control objective: y tracks yd .


I
I

I
I

First differentiations of y linear I/O relation


The control law u = x23 e(t) + y d (t) exp. convergence of e :
e + e = 0
Internal dynamics: x 2 + x23 = y d e
Since e and y d are bounded (y d (t) e D) x2 is ultimately bounded.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

15/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

I/O linearization can also be applied to stabilization (yd (t) 0):


I

I
I

For previous example the objective will be y and y will be driven to zero
and stable internal dynamics guarantee stability of the whole system.
No restriction to choose physically meaningful h(x) in y = h(x)
Different choices of output function leads to different internal dynamics
which some of them may be unstable.

When the relative degree of a system is the same as its order:


I
I

There is no internal dynamics


The problem will be input-state linearization

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

16/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Summary
I

Feedback linearization cancels the nonlinearities in a nonlinear system s.t.


the closed-loop dynamics is in a linear form.

Canceling the nonlinearities and imposing a desired linear dynamics, can


be applied to a class of nonlinear systems, named companion form, or
controllability canonical form.
When the nonlinear dynamics is not in a controllability canonical form,
input-state linearization technique is employed:

1. Transform input and state into companion canonical form


2. Use standard linear techniques to design controller
I

For tracking a desired traj, when y is not directly related to u, I/O


linearizaton is applied:
1. Generating a linear input-output relation (take derivative of y r n times)
2. Formulating a controller based on linear control

Relative degree: # of differentiating y to find explicate relation to u.

If r 6= n, there are n r internal dynamics that their stability be checked.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

17/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Internal Dynamics of Linear Systems


I

In general, directly determining the stability of the internal dynamics is


not easy since it is nonlinear. nonautonomous, and coupled to the
external closed-loop dynamics.

We are seeking to translate the concept of internal dynamics to the more


familiar context of linear systems.

Example: Consider the controllable, observable system






x2 + u
x 1
=
u
x 2
y = x1

(3)

Control objective: y tracks yd .


I
I
I
I

First differentiations of y y = x2 + u
The control law u = x2 e(t) + y d (t) exp. convergence of e : e + e = 0
Internal dynamics: x 2 + x2 = y d e
e and y d are bounded
x2 and therefore u are bounded.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

18/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Now consider a little different dynamics






x 1
x2 + u
=
x 2
u
y = x1

(4)

using the same control law yields the following internal dynamics
x 2 x2 = e(t) y d

I
I

Although yd and y are bounded, x2 and u diverge to as t


why the same tracking design method yields different results?
I
I
I
I

Transfer function of (3) is: W1 (s) = s+1


s2 .
Transfer function of (4) is: W2 (s) = s1
s2 .
Both have the same poles but different zeros
The system W1 which is minimum-phase tracks the desired trajectory
perfectly.
The system W2 which is nonminimum-phase requires infinite effort for
tracking.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

19/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Internal Dynamics
I

Consider a third-order linear system with one zero


x = Ax + bu, y = c T x
b0 +b1 s
u
a0 +a1 s+a2 s 2 +s 3

Its transfer function is: y =

First transform it into the companion

0
1
z 1
z 2 = 0
0
z 3
a0 a1

z1
y = [b0 b1 0] z2
z3

Farzaneh Abdollahi

(5)

Nonlinear Control

form:

0
0
z1

1
z2 + 0 u
z3
1
a2

(6)

Lecture 9

20/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

In second derivation of y , u appears:


y = b0 z3 + b1 (a0 z1 a1 z2 a2 z3 + u)
Required number of differentiations (the relative degree) is indeed the
same as # of poles- # of zeros
I

Note that: the I/O relation is independent of the choice of state variables
two differentiations is required for u to appear if we use (5).
b0
b1 z3 )

1
b1 (k1 e

k2 e + yd )

The control law: u = (a0 z1 + a1 z2 + a2 z3

an exp. stable tracking is guaranteed


The internal dynamics can be described by only one state equation

I
I
I

z1 can complete the state vector,( z1 , y , and y are related to z1 , z2 and z3


through a one-to-one transformation).
z1 = z2 = b11 (y b0 z1 )
y is bounded
stability of the internal dynamics depends on bb10
If the system is minimum phase the internal dynamics is stable
(independent of initial conditions and magnitude of desired trajectory)

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

21/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Zero-Dynamics
I

For linear systems the stability of the internal dynamics is determined by


the locations of the zeros.

To extend the results for nonlinear systems the concept of zero should be
modified.
Extending the notion of zeros to nonlinear systems is not trivial

In linear systems I/O relation is described by transfer function which zeros


and poles are its fundamental components. But in nonlinear systems we
cannot define transfer function
Zeros are intrinsic properties of a linear plant. But for nonlinear systems
the stability of the internal dynamics may depend on the specific control
input.

Zero dynamics: is defined to be the internal dynamics of the system when


the system output is kept at zero by the input.(output and all of its
derivatives)

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

22/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

For dynamics (2), the zero dynamics is x 2 + x23 = 0


I we find input u to maintain the system output at zero uniquely (keep x
1
zero in this example),
I By Layap. Fcn V = x 2 it can be shown it is a.s
2

For linear system (5), the zero dynamics is z 1 + (b0 /b1 )z1 = 0

The poles of the zero-dynamics are exactly the zeros of the system.

In linear systems, if all zeros are in LHP


internal dynamics.

In nonlinear systems, no results on the global stability


I only local stability is guaranteed for the internal dynamics even if the
zero-dynamics is g.e.s.

Zero-dynamics is an intrinsic feature of a nonlinear system, which does not


depend on the choice of control law or the desired trajectories.

Examining the stability of zero-dynamics is easier than examining the stability of


internal dynamics,But the result is local.
I Zero-dynamics only involves the internal states
I Internal dynamics is coupled to the external dynamics and desired trajs.

Farzaneh Abdollahi

Nonlinear Control

g.a.s. of the zero-dynamics

Lecture 9

g.s. of

23/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Zero-Dynamics
I

Similar to the linear case, a nonlinear system whose zero dynamics is


asymptotically stable is called an asymptotically minimum phase system,

If the zero-dyiamics is unstable, different control strategies should be


sought
As summary control design based on input-output linearization is in three
steps:

1. Differentiate the output y until the input u appears


2. Choose u to cancel the nonlinearities and guarantee tracking convergence
3. Study the stability of the internal dynamics
I

If the relative degree associated with the input-output linearization is the


same as the order of the system
the nonlinear system is fully linearized
satisfactory controller

Otherwise, the nonlinear system is only partly linearized


whether or
not the controller is applicable depends on the stability of the internal
dynamics.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

24/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Preliminary Mathematics
I

Vector function f : R n R n is called a vector field in R n .

Smooth vector field: function f(x) has continuous partial derivatives of


any required order.

Gradient of a smooth scalar function h(x) is denoted by a row vector


h
h = h
x , where (h)j = xj

Jacobian of a vector field f(x):an n n matrix f =


fi
(f)ij = x
j

Lie derivative of h with respect to f is a scalar function defined by


Lf h = hf, where h : R n R: a smooth scalar, f : R n R n : a smooth
vector field.

If g is another vector field: Lg Lf h = (Lf h)g

i1
L0f h = h; Lif h = Lf (Li1
f h) = (Lf h)f

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

f
x ,

where

25/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example: For single output system x = f(x), y = h(x) then


h
x = Lf h
y =
x
[Lf h]
y =
x = L2f h
x

If V is a Lyap. fcn candidate, its derivative V can be written as Lf V .

Lie bracket of f and g is a third vector field defined by


[f, g] = gf f g, where f and g two vector field on R n .

The Lie bracket [f, g] is also written as adf g (ad stands for adjoint).

adf0 g = g ; adfi g = [f , adfi1 g ], i = 1, ...

Example:
Consider x = f(x)

 + g(x)u
 where 
2x1 + ax2 + sin x1
0
f=
,g =
x2 cos x1
cos(2x1 )
So the Lie
 bracket is:

a cos(2x1 )
[f, g] =
cos x1 cos(2x1 ) 2 sin(2x1 )(2x1 + ax2 + sin x1 )

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

26/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Lemma: Lie brackets have the following properties:


1. bilinearity:
[1 f1 + 2 f2 , g]

= 1 [f1 , g] + 2 [f2 , g]

[f, 1 g1 + 2 g2 ]

= 1 [f, g1 ] + 2 [f, g2 ]

where f, f1 , f2 , g g1 , g2 are smooth vector fields and 1 and 2 are


constant scalars.
2. skew-commutativity:
[f, g]

= [g, f]

3. Jacobi identity
Ladf g h

= Lf Lg h Lg Lf h

where h is a smooth fcn.


Farzaneh Abdollahi

Nonlinear Control

Lecture 9

27/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Diffeomorphism
I

The concept of diffeomorphism can be applied to transform a nonlinear


system into another nonlinear system in terms of a new set of states.

Definition: A function : Rn Rn defined in a region is called a


diffeomorphism if it is smooth, and if its inverse 1 exists and is smooth.

If the region is the whole space Rn

Global diffeomorphisms are rare,we are looking for local diffeomorphisms.

Lemma: Let (x) be a smooth function defined in a region in Rn . If


the Jacobian matrix is non-singular at a point x = x0 of , then (x)
defines a local diffeomorphism in a subregion of

Farzaneh Abdollahi

Nonlinear Control

(x) is global diffeomorphism

Lecture 9

28/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Diffeomorphism
I

Consider the dynamic system described by


x = f (x) + g (x)u, y = h(x)

Let the new set of states z = (x)

The new state-space representation


z = f (z) + g (z)u, y = h (z)

z =

x x

x (f (x)

+ g (x)u)

where x = 1 (z).
Example
diffeomorphism:
Consider

 of a non-global


z1
2x1 + 5x1 x22
= (x) =
z2
3 sin x2
I
I


2 + 5x22 10x1 x2
=
Its Jacobian matrix:
.
0
3 cos x2
rank is 2 at x = (0, 0) local diffeomorphism around the origin where
= {(x1 , x2 ), |x2 | < /2}.
outside the region, the inverse of does not uniquely exist.

I
Farzaneh Abdollahi

Nonlinear Control

Lecture 9

29/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Frobenius Theorem
I

An important tool in feedback linearization

Provide necess. and suff. conditions for solvability of PDEs.

Consider a PDE with (n=3):


h
h
h
f1 +
f2 +
f3 = 0
x1
x2
x3
h
h
h
g1 +
g2 +
g3 = 0
x1
x2
x3

(7)

where fi (x1 , x2 , x3 ), gi (x1 , x2 , x3 ) (i = 1, 2, 3) are known scalar fcns and


h(x1 , x2 , x3 ) is an unknown function.
I

This set of PDEs is uniquely determined by the two vectors


f = [f1 f2 f3 ]T , g = [g1 g2 g3 ]T .

If the solution h(x1 , x2 , x3 ) exists, the set of vector fields {f , g } is


completely integrable.

When the equations are solvable?

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

30/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Frobenius Theorem
I

Frobenius theorem states that Equation (7) has a solution h(x1 , x2 , x3 ) iff
there exists scalar functions 1 (x1 , x2 , x3 ) and 2 (x1 , x2 , x3 ) such that
[f , g ] = 1 f + 2 g
i.e., if the Lie bracket of f and g can be expressed as a linear combination of f
and g

This condition is called involutivity of the vector fields {f , g }.

Geometrically, it means that the vector field [f , g ] is in the plane formed by the
two vectors f and g

The set of vector fields {f , g } is completely integrable iff it is involutive.

Definition (Complete Integrability): A linearly independent set of vector


fields {f1 , f2 , ..., fm } on R n is said to be completely integrable, iff, there exist
n m scalar fcns h1 (x), h2 (x), ..., hnm (x) satisfying the system of PDEs:
hi fj = 0
where 1 i n m, 1 j m and hi are linearly independent.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

31/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Number of vectors: m, dimension of the vectors: n, number of unknown


scalar fcns hi : (n-m), number of PDEs: m(n-m)

Definition (Involutivity): A linearly independent set of vector fields


{f1 , f2 , ..., fm } on R n is said to be involutive iff, there exist scalar fcns
ijk : R N R s.t.
m
X
[fi , fj ](x) =
ijk (x) fk (x)
i, j
k=i

i.e., the Lie bracket of any two vector fields from the set {f1 , f2 , ..., fm }
can be expressed as the linear combination of the vectors from the set.
I
I

Constant vector fields are involutive since their Lie brackets are zero
A set composed of a single vector is involutive:
[f , f ] = (f )f (f )f = 0
Involutivity means:
rank (f1 (x) .... fm (x)) = rank (f1 (x) .... fm (x) [fi , fj ](x))
for all x and for all i, j.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

32/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Frobenius Theorem
I

Theorem (Frobenius): Let f1 , f2 , ...., fm be a set of linearly


independent vector fields. The set is completely integrable iff it is
involutive.

Example: Consider the set of PDEs:


h
h

x1 x2
h
h
h
x1
+ (x32 3x2 )
+ 2x3
x1
x2
x3
4x3

= 0
= 0

The associated vector fields are {f1 , f2 }


f1 = [4x3 1 0]T

f2 = [x1 (x32 3x2 ) 2x3 ]T

We have [f1 , f2 ] = [12x3 3 0]T

Since [f1 , f2 ] = 3f1 + 0f2 , the set {f1 , f2 } is involutive and the set of
PDEs are solvable.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

33/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-State Linearization
I

Consider the following SISO nonlinear system


x = f (x) + g (x)u

(8)

where f and g are smooth vector fields


I

The above system is also called linear in control or affine

If we deal with the following class of systems:


x = f (x) + g (x)w (u + (x))
where w is an invertible scalar fcn and is an arbitrary fcn
I We can use v = w (u + (x)) to get the form (8).
I Control design is based on v and u can be obtained by inverting w :
u = w 1 (v ) (x)

Now we are looking for


I Conditions for system linearizability by an input-state transformation
I A technique to find such transformations
I A method to design a controller based on such linearization technique

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

34/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-State Linearization
I

Definition: Input-State Linearization The nonlinear system (8) where f (x)


and g (x) are smooth vector fields in R n is input-state linearizable if there exist
region in R n , a diffeomorphism mapping : R n , and a control law:
u = (x) + (x)v
s.t. new state variable z = (x) and
z = Az + Bv

0 1 0
0 0 1

. . .
A =
. . .

0 0 0
0 0 0

new input variable v satisfy an LTI relation:

... 0
... .

... .

... .

... 1
... 0

0
0
..
.

B=

0
1

(9)

The new state z is called the linearizing state and the control law u is called the
linearizing control law

Let z = z(x)

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

35/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-State Linearization
I

(9) is the so-called linear controllability or companion form

This linear companion form can be obtained from any linear controllable
system by a transformation
if u leads to a linear system, (9) can be
obtained by another transformation easily.

This form is an special case of Input-Output linearization leading to


relative degree r = n.

Hence, if the system I/O linearizable with r = n, it is also I/S linearizable.

On the other hand, if the system is I/S linearizable, it is also I/O


linearizable with y = z, r = n.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

36/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-State Linearization
I

Lemma: An nth order nonlinear system is I/S linearizable iff there exists
a scalar fcn z1 (x) for which the system is I/O linearizable with r = n.

Still no guidance on how to find the z1 (x).


Conditions for Input-State Linearization:

I
I

Theorem: The nonlinear system (8) with f (x) and g (x) being smooth
vector field is input-state linearizable iff there exists a region s.t. the
following conditions hold:
The vector fields {g , adf g , ... adf n1 g } are linearly independent in
The set {g , adf g , ... adf n2 g } is involutive in

The first condition:


I
I
I

can be interpreted as a controllability condition


For linear system, the vector field above becomes {B, AB, ... An1 B}
Linear independency invertibility of controllability matrix

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

37/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

The second condition


I

is always satisfied for linear systems since the vector fields are constant, but
for nonlinear system is not necessarily satisfied.
It is necessary according to Ferobenius theorem for existence of z1 (x).

Lemma: If z(x) is a smooth vector field in , then the set of equations


Lg z = Lg Lf z = ... = Lg Lf k z = 0
is equivalent to
Lg z = Ladf

gz

= ... = Ladf

kg

z =0

Proof:
I

Let k = 1, from Jacobis identity, we have


Ladf g z = Lf Lg z Lg Lf z = 0 0 = 0

When k = 2, we have from Jacobis identity:


Ladf

Farzaneh Abdollahi

2g

z = Lf 2 Lg z 2Lf Lg Lf z + Lg Lf 2 z = 0 0 + 0 = 0
Nonlinear Control

Lecture 9

38/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

I
I

Proof of the linearization theorem:


Necessity:
I

Suppose state transformation z = z(x) and input transformation


u = (x) + (x)v s.t. z and v satisfy (9), i.e.
z1
(f + gu) = z2
z 1 =
x
similarly:
z1
z1
f +
gu
x
x
z2
z2
f +
gu
x
x

= z2
= z3
..
.

zn
zn
f +
gu
x
x

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

= v

39/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

z1 , ..., zn1 are independent of u,


Lg z1 = Lg z2 = ... Lg zn1 = 0, Lg zn 6= 0
Lf zi

= zi+1 , i = 1, 2, ..., n 1

Use, z = [z1 Lf z1 , ... Lf n1 z1 ]T to get


z k = zk+1 , k = 1, ... n 1
z n = Lf n z1 + Lg Lf

n1

z1 u

The above equations can be expressed in terms of z1 only


z1 adf k g = 0, k = 0, 1, 2, ..., n 2
z1 adf

n1

n1

= (1)

Lg zn

First note that for above eqs to hold, the vector field g , adf g , ..., adf
must be linearly independent.

If for some i(i n 1) there exist scalar fcns 1 (x), ... i1 (x) s.t.
i1
X
adf i g =
k adf k g

(10)
(11)
n1

k=0
Farzaneh Abdollahi

Nonlinear Control

Lecture 9

40/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

We, then have:


adf

n1

n2
X

k adf

k=ni1

z1 adf

n1

n2
X

k z1 adf

g =0

(12)

k=ni1

Contradicts with (11).


I

The second property is that a scalar fcn z1 that satisfy n 1 PDEs


z1 adf k g = 0

From the necessity part of Frobenius theorem, we conclude that the set
of vector field must be involutive.
Sufficient condition

Involutivity condition = Frobenius theorem, a scalar fcn z1 (x):


Lg z1 = Ladf g z1 = ... Ladf k g z1 = 0, implying
Lg z1

Farzaneh Abdollahi

= Lg Lf z1 = ... Lg Lf
Nonlinear Control

Lecture 9

z1 = 0
41/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Define the new sets of variable as z = [z1 Lf z1 ... Lf n1 z1 ]T , to get


z k = zk+1
k = 1, ..., n 1
z n = Lf n z1 + Lg Lf
The question is whether Lg Lf

n1 z
1

n1

z1 u

(13)

can be equal to zero.

n1 g }

Since {g , adf g , ..., adf


are linearly independent in :
n1
z1 = (1)n1 Ladf n1 g z1
Lg Lf

We must have Ladf n1 g z1 6= 0, otherwise the nonzero vector z1


satisfies
z1 [g , adf g , ..., adf n1 g ] = 0
i.e. z1 is normal to n linearly independent vector = impossible

Now, we have:
z n = Lf n z1 + Lg Lfn1 z1 u = a(x) + b(x)u

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

42/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

1
b(x) (a(x)

Now, select u =

+ v ) to get:
z n = v

implying input-state linearization is obtained.


Summary: how to perform input-state Linearization
1. Construct the vector fields g , adf g , ... adf n1 g
2. Check the controllability and involutivity conditions
3. If the conditions hold, obtain the first state z1 from:
z1 adf i g = 0 i = 0, ..., n 2
z1 adf

n1

6=

4. Compute the state transformation z(x) = [z1 Lf z1 ... Lf


input transformation u = (x) + (x)v :
Lf n z1
(x) =
Lg Lf n1 z1
1
(x) =
Lg Lf n1 z1
Farzaneh Abdollahi

Nonlinear Control

Lecture 9

n1

z1 ]T and the

43/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example:A single-link flexible-joint manipulator:


I

The link is connected to the motor shaft via a torsional spring

Equations of motion:
I q1 + MgLsinq1 + K (q1 q2 ) = 0
J q2 K (q1 q2 ) = u

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

44/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example:A single-link flexible-joint manipulator:


I

Equations of motion:

I q1 + MgLsinq1 + K (q1 q2 ) = 0
J q2 K (q1 q2 ) = u

nonlinearities appear in the first equation and torque is in the second equation

Let:

x2
q1
q 1
MgL sinx1 K (x1 x3 )

I
I
x =
q2 , f =
x4
K
q 2
J (x1 x3 )
I

Controllability and involutivity conditions:


0
0

0
0
[g adf g adf 2 g adf 3 g ] =
0 1
J
1
0
J

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

, g = 0

0
K
IJ

0
JK2

1
J

K
IJ
0

J2
0
45/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example: Contd
I

Its full rank for k > 0 and IJ < = vector fields are linearly
independent

Vector fields are constant = involutive

The system is input-state linearizable

Computing z = z(x), u = (x) + (x)v

I z1 = 0, z1 = 0, z1 = 0, z1 6= 0
x2
x3
x4
x1
I Hence, z1 is the fcn of x1 only. Let z1

= x1 , then

z2 = z1 f = x2
MgL
K
sinx1 (x1 x3 )
I
I
MgL
K
= z3 f =
x2 cosx1 (x2 x4 )
I
I

z3 = z2 f =
z4
Farzaneh Abdollahi

Nonlinear Control

Lecture 9

46/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example: Contd
I

The input transformation is given by:


u

a(x)

=
+

IJ
(v a(x))
K
MgL
MgL
K
sinx1 (x22 +
cosx1 + )
I
I
I
K
K
K
MgL
(x1 x3 )( + +
cosx1 )
I
I
J
I

(v z4 f )/(z4 g ) =

As a result, we get the following set of linear equations


z 1 = z2 , z 2 = z3
z 3

= z4 ,

z 4 = v

The inverse of the state transformation is given by:


x1 = z1 , x2 = z2


MgL
I
x3 = z1 +
z3 +
sinz1
K
I


I
MgL
x4 = z2 +
z4 +
z2 cosz1
K
I

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

47/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

State and input transformations are defined globally

In this example, transformed state have physical meaning, z1 : link


position, z2 : link velocity, z3 : link acceleration, z4 : link jerk.

It could be obtained by I/O linearization, i.e. by differentiating the output


q1 . (4 times)

We can transform the inequality (11) to a normalized equation by setting


z1 adf n1 g = 1 resulting in:

0
z1
0
x1

.. ..
0
1
n2
n1
[adf g adf g ... adf
g adf
g ] . = .

z1
0
xn
1

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

48/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Control Design
I

Once, the linearized dynamics is obtained, either a tracking or


stabilization problem can be solved

For instance, in flexible-joint manipulator case, we have


(4)

z1
I

=v

Then, a tracking controller can be obtained as


(4)
(3)
1 a1 z 1 a0 z1
v = z a3 z a2 z
d1

where z1 = z1 zd1 .
I

The error dynamics is then given by:


(4)
(3)
z1 + a3 z1 + a2 z
1 + a1 z 1 + a0 z1 = 0

The above dynamics is exponentially stable if ai are selected s.t.


s4 + a3 s 3 + a2 s 2 + a1 s + a0 is Hurwitz

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

49/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Input-Output Linearization
I

Consider the system:


= f (x) + g (x)u

= h(x)

(14)

Input-output linearization yields a linear relationship between the output


y and the input v (similar to v in I/S Lin.)
I
I

How to generate a linear I/O relation for such systems?


What are the internal dynamics and zero-dynamics associated with this I/O
linearization
How to design a stable controller based on this technique?

Performing I/O Linearization


I

The basic approach is to differentiate the output y until the input u


appears, then design u to cancel nonlinearities
Sometime, cancelation might not be possible due to the undefined relative
degree.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

50/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Well Defined Relative Degree


I

Differentiate y and express it in the form of Lie derivative:


y = h(f + gu) = Lf h(x) + Lg h(x)u
if Lg h(x) 6= 0 for some x = x0 in x , then continuity implies that
Lg h(x) 6= 0 in some neighborhood of x0 . Then, the input
transformation
1
(Lf h(x) + v )
u=
Lg h(x)
results in a linear relationship between y and v , namely y = v .

If Lg h(x) = 0 for all x x , differentiate y to obtain


y = Lf 2 h(x) + Lg Lf h(x)u

If Lg Lf h(x) = 0 for all x x , keep differentiating until for some


integer r , Lg Lf r 1 h(x) 6= 0 for some x = x0 x

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

51/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Hence, we have
y (r ) = Lf r h(x) + Lg Lf

r 1

h(x)u

(15)

and the control law


u=

1
Lg Lf

r 1 h(x)

(Lf r h(x) + v )

yields a linear mapping:


y (r ) = v
I

The number r of differentiation required for u to appear is called the


relative degree of the system.

r n, if r = n, the input-state realization is obtained with z1 = y .

Definition: The SISO system is said to have a relative degree r in if:


Lg Lf i h(x) = 0
0 i r 2
Lg Lf

Farzaneh Abdollahi

r 1

h(x) 6= 0
Nonlinear Control

Lecture 9

52/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Undefined Relative Degree


I

Sometimes, we are interested in the properties of a system about a


specific operating point x0 .

Then, we say the system has relative degree r at x0 if


Lg Lf

r 1

h(x0 ) 6= 0
r 1 h(x)

However, it might happen that Lg Lf


a close neighborhood of x0 .

is zero at x0 , but nonzero in

The relative degree of the nonlinear system is then undefined at x0 .

Example:
x = (x, x)
+u
where is a smooth nonlinear fcn. Define x = [x x]
T and let y = x =
the system is in companion form with r = 2.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

53/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

However, if we define y = x 2 , then:


y

= 2x x

y = 2x x + 2x 2 = 2x(x, x)
+ 2xu + 2x 2 =
Lg Lf h = 2x

(16)

The system has neither relative degree 1 nor 2 at x0 = 0.

Sometime, change of output leads us to a solvable problem.

We assume that the relative degree is well defined.


Normal Forms

When, the relative degree is defined as r n, using y , y , ..., y (r 1) , we


can transform the system into the so-called normal form.
Normal form allows a formal treatment of the notion of internal dynamics
and zero dynamics.
Let
= [1 2 ... r ]T = [y y ... y (r 1) ]T
in a neighborhood of a point x0 .

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

54/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Normal Form
I

The normal form of the system can be written as

..

.
=

r
a(, ) + b(, )u
= w (, )

(17)

(18)

= 1

The i and j are called normal coordinate or normal states.

The first part of the Normal form, (17) is another form of (15), however in (18)
the input u does not appear.

The system can be transformed to this form if the state transformation (x) is a
local diffeomorphism: (1 ... r 1 ... nr )T

To show that is a diffeomorphism, we must show that the Jacobian is


invertible, i.e. i and i are all linearly independent.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

55/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Normal Form
I

i are linearly independent = can be part of state variables, ( is


output and its r 1 derivatives)

There exist n r other vector fields that complete the transformation

Note that u does not appear in (18), hence:


j g = 0 1 j n r
can be obtained by solving n r PDE above.

I
I

Generally, internal dynamics can be obtained simpler by intuition.


Zero Dynamics
I

System dynamics have two parts:


1. external dynamics

2. internal dynamics

For tracking problems (y yd ), one can easily design v once the linear
relation is obtained.
The question is whether the internal dynamics remain bounded

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

56/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Zero-Dynamics
I

Stability of the zero dynamics (i.e. internal dynamics when y is kept 0)


gives an idea about the stability of internal dynamics

u is selected s.t. y remains zero at all time.


y (r ) (t) = Lf r h(x) + Lg Lf
Lf r h(x)
u0 (x) =
Lg Lf r 1 h(x)

r 1

h(x)u0 0 =

In normal form:


=
0

= w (0, )
a(0, )
u0 () =
b(0, )

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

(19)
57/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example
I

x
y
I

We have

2x
x1
e 2
= 2x1 x2 + sinx2 + 1/2 u
0
2x2
= h(x) = x3

Consider

= 2x2

y = 2x 2 = 2(2x1 x2 + sinx2 ) + u
I

The system has relative degree r = 2 and


Lf h(x) = 2x2
Lg h(x) = 0
Lg Lf h(x) = 1

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

58/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

To obtain the normal form


1 = h(x) = x3
2 = Lf h(x) = 2x2

The third function (x) is obtained by


2x2 1
e +
=0
Lg =
x1
2 x2

One solution is (x) = 1 + x1 e 2x2

Consider the jacobian of state transformation z = [1 2 ]T . The


Jacobian matrix is

0
0
1
0
2
0
2x
2
1 2e
0

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

59/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

The Jacobian is non-singular for any x. In fact, inverse transformation is given


by:
x1
x2
x3

= 1 + + e 2
1
2
=
2
= 1

State transformation is valid globally and the normal form is given by:
1

2 = 2(1 + + e 2 )2 + 2sin(2 /2) + u


= (1 e 2 )(1 + 22 e 2 ) 2sin(2 /2)e 2

(20)

Zero dynamics is obtained by setting 1 = 2 = 0 =


=

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

(21)
60/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Zero-Dynamics
I

In order to obtain the zero dynamics, it is not necessary to put the system
into normal form

since is known, we can intuitively find n r vector to complete the


transformation.

As mention before, zero dynamics is obtained by substituting u0 for u in


internal dynamics.

Definition: A nonlinear system with asymptotically stable zero dynamics


is called asymptotically minimum phase

If the zero dynamics is stable for all x, the system is globally minimum
phase, otherwise the results are local.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

61/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Local Asymptotic Stabilization


I

Consider again the nonlinear system


x = f (x) + g (x)u
y

h(x)

Assume that the system is I/O linearized, i.e.


y (r ) = Lf r h(x) + Lg Lf r 1 h(x)u
and the control law
u=
yields a linear mapping:
I

1
Lg Lf

r 1 h(x)

(Lf r h(x) + v )

(22)

(23)
(24)

y (r ) = v

Now let v be chosen as


v = kr 1 y (r 1) ... k1 y k0 y

(25)

where ki are selected s.t. K (s) = s r + kr 1 s r 1 + ... + k1 s + k0 is Hurwitz


Farzaneh Abdollahi

Nonlinear Control

Lecture 9

62/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Then, provided that the zero-dynamics is asymptotically stable, the control law
(24) and (25) locally stabilize the whole system:

Theorem: Suppose the nonlinear system (22) has a well defined relative degree
r and its associated zero-dynamics is locally asymptotically stable. Now, if ki are
selected s.t. K (s) = s r + kr 1 s r 1 + ... + k1 s + k0 is Hurwitz, then the control
law (24) and (25) yields a locally asymptotically stable system.

Proof: First, write theclosed-loop system in a normal form:

0
1
0
...
0
0

0
1
...
.

.
.
...
.
= A
=
.

.
.
...
.

0
0
...
1
k0 k1 k2 . . . kr 1

= w (, ) = A1 + A2 + h.o.t.
h.o.t. is higher order terms in the Taylor expansion about x0 = 0.
The above Eq. can be written
as:

 

d

A
0

=
+ h.o.t.
A1 A2

dt

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

63/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

= A2
Now, since the zero dynamics is asymptotically stable, its linearization
is either asymptotically stable or marginally stable.
I If A is asymptotically stable, then all eigenvalues of the above system
2
matrix are in LHP and the linearized system is stable and the nonlinear
system is locally asymptotically stable
I If A is marginally stable, asymptotic stability of the closed-loop system
2
was shown in (Byrnes and Isidori, 1988).

Comparing the above method to local stabilization and using linear control:
I the above stabilization method can treat systems whose linearizations
contain uncontrollable but marginally stable modes,
I while linear control methods requires the linearized system to be strictly
stabilizable

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

64/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

For stabilization where state convergence is required, we can freely


choose y = h(x) to make zero-dynamics a.s.

Example: Consider the nonlinear system:


x 1 = x12 x2
x 2 = 3x2 + u

System linearization at x = 0:
x 1 = 0
x 2 = 3x2 + u
thus has an uncontrollable mode

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

65/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example (contd)
I Define y = 2x x
1
2 =
y
I

2x 1 x 2 = 2x12 x2 3x2 u

Hence, the relative degree r = 1 and the associated zero-dynamics is


x 1

2x13

The zero-dynamics is asymptotically stable, hence the control law


u = 2x12 x2 4x2 2x1 locally stabilizes the system

Global Asymptotic Stabilization

Stability of the zero-dynamics only guarantees local stability unless relative


degree is n in which case there is no internal dynamics

Can the idea of I/O linearization be used for global stabilization problem?

Can the idea of I/O linearization be used for systems with unstable zero
dynamics?

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

66/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Global Asymptotic Stabilization


I

Global stabilization approach based on partial feedback linearization is to


simply regard the problem as a standard Lyapunov controller design
problem

But simplified by the fact that in normal form part of the system
dynamics is now linear.
The basic idea is to view as the input to the internal dynamics and
as its output.

The first step: find the control law 0 = 0 () which stabilizes the internal
dynamics with the corresponding Lyapunov fcn V0 .
Then: find a Lyapunov fcn candidate for the whole system (as a modified
version of V0 ) and choose the control input v s.t. V be a Lyapunov fcn for
the whole closed-loop dynamics.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

67/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example:
I

Consider a nonlinear system with the normal form:


y

= v

z + z 3 + yz = 0

(26)

where v is the control input and = [z z]


T
Considering y as an input to internal dynamics (26), it would be
asymptotically stabilized by the choice of y = y0 = z 2
I

Let V0 be a Lyap. fcn:


V0 =

1 2 1 4
z + z
2
4

Differentiating V0 along the actual dynamics results in


V 0

Farzaneh Abdollahi

= z 4 z z(y
z 2)

Nonlinear Control

Lecture 9

68/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

Consider the Lyap. fcn candidate, obtained by adding a quadratic error


term in y y0 to V0
V
V

1
= V0 + (y z 2 )2
2
= z 4 + (y z 2 )(v 3z z)

The following choice of control action will then make V n.d.


v = y + z 2 + 3z z
V = z 4 (y z 2 )2

Application of Invariant-set theorem shows all states converges to zero

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

69/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example: A non-minimum phase system


I

Consider the system dynamics


y = v
z + z 3 z 5 + yz = 0
where again = [z z]
T
I The system is non-minimum phase since its zero-dynamics is unstable
I The zero-dynamics would be stable if we select y = 2z 4 :
1 2 1 6
V0 =
z + z
V0 = z 4 z z(y
2z 4 )
2
6
I

Consider the Lyap. fcn candidate


1
V = V0 + (y 2z 4 )2 V = z 4 + (y 2z 4 )(v 8z 3 z z z)

suggesting the following choice of control law


v = y + 2z 4 + 8z 3 z + z z V = z 4 (y 2z 4 )2

Application of Invariant-set theorem shows all states converges to zero

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

70/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Tracking Control
I

I/O linearization can be used in tracking problem

Let d = [yd y d ... yd

Theorem: Assume the system (22) has a well defined relative degree r
and d is smooth and bounded and that the solution d :
d = w (d , d ), d (0) = 0

(r 1) T
]

and the tracking error


(t) = (t) d (t)

exists and bounded and is uniformly asymptotically stable. Choose ki s.t


K (s) = s r + kr 1 s r 1 + ... + k1 s + k0 is Hurwitz, then by using
1
(r )
[Lf r 1 + yd kr 1
r ... k0
1 ]
(27)
u=
r
1
Lg Lf
1
the whole system remains bounded and the tracking error
converge to
zero exponentially.
I

Proof: Refer to Isidori (1989).

For perfect tracking (0) d (0)

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

71/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Tracking Control for Non-minimum Phase Systems:


I

The tracking control (27) cannot be applied to non-minimum phase


systems since they cannot be inverted

Hence we cannot have perfect or asymptotic tracking and should seek


controllers that yields small tracking errors
One approach is the so-called Output redefinition

I
I

The new output y1 is defined s.t. the associated zero-dynamics is stable


y1 is defined s.t. it is close to the original output y in the frequency range
of interest
Then, tracking y1 also implies good tracking the original output y

Example: Consider a linear system



1 bs B0 (s)
u
y=
A(s)

b>0

Perfect/asymptotic tracking is impossible due to the presence of zero @


s=b

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

72/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

Let us redefine the output as


y1 =

B0 (s)
u
A(s)

with the desired output for y1 be simply yd


I

A controller can be found s.t. y1 asymptotically tracks yd . What about


the actual tracking error?

s
s
y (s) = 1
y1 = 1
yd
b
b

Thus, the tracking error is proportional to the desired velocity y d :


y d (t)
y (t) yd (t) =
b

Tracking error is bounded as long as y d is bounded, it is small when


the frequency content of yd is well below b

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

73/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Example Contd
I

An alternative output, motivated by (1 bs ) 1/(1 + bs ) for small |s|/b:


y2 =

B0 (s)
u
A(s)(1 + bs )



s 
s
s2
y (s) = 1
1+
yd = 1 2 yd
b
b
b


Thus, the tracking error is proportional to the desired acceleration yd :


y (t) yd (t) =

yd (t)
b2

Small tracking error if the frequency content of yd is below b

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

74/75

Outline Feedback Linearzation Preliminary Mathematics Input-State Linearization Input-Output Linearization

Tracking Control
I

Another approximate tracking (Hauser, 1989) can be obtained by


I

I
I
I

When performing I/O linearization, using successive differentiation, simply


neglect the terms containing the input
Keep differentiating n timed (system order)
Approximately, there is no zero dynamics
It is meaningful if the coefficients of u at the intermediate steps are small
or the system is weakly non-minimum phase system
The approach is similar to neglecting fast RHP zeros in linear systems.

Zero-dynamics is the property of the plant, choice of input and output


and desired Trajectory. It cannot be changed by feedback:
I

I
I
I

Modify the plant (distribution of control surface on an aircraft or the mass


and stiffness in a flexible robot)
Change the output (or the location of sensor)
Change the input (or the location of actuator)
Change the desired Traj.

Farzaneh Abdollahi

Nonlinear Control

Lecture 9

75/75